MMNMT: Modularizing Multilingual Neural Machine Translation with Flexibly Assembled MoE and Dense Blocks

Shangjie Li, Xiangpeng Wei, Shaolin Zhu, Jun Xie, Baosong Yang, Deyi Xiong


Abstract
Mixture-of-Experts (MoE) based sparse architectures can significantly increase model capacity with sublinear computational overhead, which are hence widely used in massively multilingual neural machine translation (MNMT). However, they are prone to overfitting on low-resource language translation. In this paper, we propose a modularized MNMT framework that is able to flexibly assemble dense and MoE-based sparse modules to achieve the best of both worlds. The training strategy of the modularized MNMT framework consists of three stages: (1) Pre-training basic MNMT models with different training objectives or model structures, (2) Initializing modules of the framework with pre-trained couterparts (e.g., encoder, decoder and embedding layers) from the basic models and (3) Fine-tuning the modularized MNMT framework to fit modules from different models together. We pre-train three basic MNMT models from scratch: a dense model, an MoE-based sparse model and a new MoE model, termed as MoE-LGR that explores multiple Language-Group-specifc Routers to incorporate language group knowledge into MNMT. The strengths of these pre-trained models are either on low-resource language translation, high-resource language translation or zero-shot translation. Our modularized MNMT framework attempts to incorporate these advantages into a single model with reasonable initialization and fine-tuning. Experiments on widely-used benchmark datasets demonstrate that the proposed modularized MNMT framwork substantially outperforms both MoE and dense models on high- and low-resource language translation as well as zero-shot translation. Our framework facilitates the combination of different methods with their own strengths and recycling off-the-shelf models for multilingual neural machine translation. Codes are available at https://github.com/lishangjie1/MMNMT.
Anthology ID:
2023.emnlp-main.303
Volume:
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4978–4990
Language:
URL:
https://aclanthology.org/2023.emnlp-main.303
DOI:
10.18653/v1/2023.emnlp-main.303
Bibkey:
Cite (ACL):
Shangjie Li, Xiangpeng Wei, Shaolin Zhu, Jun Xie, Baosong Yang, and Deyi Xiong. 2023. MMNMT: Modularizing Multilingual Neural Machine Translation with Flexibly Assembled MoE and Dense Blocks. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 4978–4990, Singapore. Association for Computational Linguistics.
Cite (Informal):
MMNMT: Modularizing Multilingual Neural Machine Translation with Flexibly Assembled MoE and Dense Blocks (Li et al., EMNLP 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.emnlp-main.303.pdf
Video:
 https://aclanthology.org/2023.emnlp-main.303.mp4